BACKGROUND: Although voriconazole reached the bedside 10 years ago and became the standard care in the treatment of invasive aspergillosis, reliable clinical breakpoints are still in high demand. Moreover, this has increased due to the recent emergence of azole resistance. METHODS: Four clinical wild-type and non-wild-type A. fumigatus isolates with voriconazole CLSI MICs in the range of 0.125-2 mg/L were tested in an in vitro pharmacokinetic (PK)/pharmacodynamic (PD) model. Mouse PK was simulated and in vitro data were compared with in vivo outcome. Human PK was simulated and susceptibility breakpoints and trough levels required for optimal treatment were determined for the CLSI and EUCAST methods after 48 h and the gradient concentration MIC test strip (MTS) method after 24 h using the in vitro PK/PD relationship and Monte Carlo simulation. RESULTS: The in vitro PK/PD target (95% CI) associated with 50% of the maximal antifungal activity (EC50) was 28.61 (16.18-50.61), close to the in vivo EC50 of 14.67 (9.31-21.58) fAUC0-24/CLSI MIC. When human PK was simulated, the EC50 was 24.7 (17.9-35.6) fAUC0-12/CLSI MIC and it was associated with 6 week survival in clinical studies of invasive pulmonary aspergillosis. Target attainment rates were =5% (0%-24%), 42% (16%-58%), 68% (54%-75%) and >/=79% (73%-86%) for isolates with CLSI MICs >/=2, 1, 0.5 and =0.25 mg/L, respectively. A trough/CLSI MIC ratio of 2 was required for optimal treatment. The susceptible/intermediate/resistant breakpoints were determined to be 0.25/0.5-1/2 mg/L for CLSI, 0.5/1-2/4 mg/L for EUCAST and 0.25/0.375-1/1.5 mg/L for MTS. CONCLUSIONS: These susceptibility breakpoints and target values for therapeutic drug monitoring could be used to optimize voriconazole therapy against A. fumigatus.